Senior Principal AI Scientist, Generative Models for Mobility

Cerence AI; United States; Hybrid; USD 185,000 - 280,000; Full time

Benefits offered by this job

Annual bonus

Insurance coverage

Paid time off

Paid holidays

RRSP contributions

Equity awards

Remote and/or hybrid work

Job summary

Cerence Inc. is hiring a Senior Principal AI Scientist in Generative AI to lead the design, training, and evaluation of large-scale models for automotive AI and mobility experiences.

You will own model architectures across text and multimodal tasks, driving stability and efficiency in distributed training environments. The role emphasizes deep theoretical and practical ML expertise, with collaboration across ML systems and product teams to scale cutting-edge models in a dynamic, global

Qualifications

• Strong theoretical and practical understanding of modern deep learning (transformers, representations).
• Hands-on experience training large models from scratch.
• Ability to reason about optimization and training dynamics beyond hyperparameter tuning.
• Comfort operating in research-driven, ambiguous environments.

Responsibilities

• Design and train large-scale transformer and foundation models for Generative AI.
• Own architecture choices across text, multimodal, and new paradigms.
• Diagnose training instabilities and optimize scaling across data and compute.
• Collaborate with ML systems teams to deploy scalable training pipelines.

Skills

Deep learning theory

Hands-on model training

Optimization reasoning

Ambiguity tolerance

Tools

FSDP

ZeRO-3

Tensor parallelism

Job description

Cerence Inc. is hiring a Senior Principal AI Scientist in Generative AI to lead the design, training, and evaluation of large-scale models for automotive AI and mobility experiences.

You will own model architectures across text and multimodal tasks, driving stability and efficiency in distributed training environments. The role emphasizes deep theoretical and practical ML expertise, with collaboration across ML systems and product teams to scale cutting-edge models in a dynamic, global